On Tue, 16 Feb 1999 21:54:02 +0100 
 Chuck Cranor <chuck@research.att.com> wrote:

 > as i understand it the problem comes about when you've got a seperate
 > I and D cache.   in that case you need to flush the i-cache properly
 > when ptrace is used.

Right.

 > the ptrace man page has a PT_WRITE_I and PT_READ_I directive, but
 > internally the kernel treats these two exactly the same (which seems
 > wrong to me).  

Right.

 > i had the impression that there should be some sort MD hook to flush
 > the I-cache in the PT_WRITE_I case?

Right.

:-)

I don't think uvm_io() should do this automagically, so that callers of
uvm_io() can optimize the cache flush calls.

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>